The Case for Observation-Calibrated Climate Data
Weather and climate profoundly influence architecture, engineering, and energy system design. However, the utility of weather and climate data depends on its accuracy. Historically, in-situ observations from weather stations have provided the most reliable reference for climate baselines. These ground-based observations remain the gold standard, offering the closest approximation to “ground truth.”
However, in recent years, many practitioners have turned to gridded weather datasets like ERA5, which are synthesized model outputs with complete global coverage, hourly resolution, and are easily accessible via cloud platforms. These features have made ERA5 immensely popular for climate risk applications, including architecture, engineering, and energy system design applications.
But here’s the problem:
ERA5 is not always direct substitute for observations.
It can reproduce patterns that look realistic but the actual values themselves and their variability can diverge significantly from measured conditions. And that distinction can have material consequences for applications that demand high levels of accuracy.
A Simple Example: Hot Days
To illustrate the differences between gridded weather datasets and direct observations, we compared two widely used gridded climate datasets against in-situ station data from HadISD:
- ERA5 – A leading global reanalysis dataset produced by ECMWF
- Global Meteorological Forcing Dataset (GMFD) – A hybrid product combining surface observations with NCEP reanalysis
GMFD is not as popular as ERA5, but it is the historical reference used by NASA’s NEX-GDDP, one of the most commonly used downscaled climate datasets among insurance companies and climate risk and resilience consultants.
As a simple example, we examined the number of days per year exceeding 90°F across thousands of weather station locations within each dataset.
At first glance, ERA5 and GMFD appear to reasonably match the observational data. However, a closer look reveals significant differences. In many regions, especially in coastal zones and tropical areas, both gridded datasets exhibit substantial bias relative to in-situ observations. In some cases, these datasets over- or under-estimate 90°F days by up to 200 days per year.
Many practitioners rely on ERA5 or GMFD as if they were interchangeable with weather station observations, using them to guide engineering design, asset-level risk assessments, or resilience planning. But this reliance can be misplaced for applications requiring local accuracy. Gridded datasets were not designed to represent conditions at a single point, but rather to provide area-averaged estimates over regions. As such, they may fail to capture key local features, including urban heat islands, coastal gradients, or microclimate effects that matter for infrastructure and energy systems.
Why Accuracy Matters
There are material reasons why the fidelity of weather and climate data is critical for engineering design, risk analysis, and other applications that depend on accurate environmental inputs.
1. Biased Baselines Lead to Biased Future Climate Projections
ERA5 and GMFD are commonly used to calibrate climate model projections. But if these foundational datasets are biased, then any future climate projection derived from them will carry forward those errors, potentially resulting in under-designed infrastructure, underestimated exposure, or misguided adaptation strategies.
It’s like a pharmacy using a broken scale to dose medicine: all of its customers will get the wrong number of pills. The same logic applies to future climate projections calibrated to inaccurate data.
2. Small Biases Can Have Big Consequences
Disruptions to infrastructure systems can have non-linear relationships with temperature. For example, a +1 °C bias in baseline temperature could translate into large changes in energy demand, building load profiles, material degradation rates, or even public health risks. This can cascade into significant miscalculations, opening the door to increased costs, performance failures, or liability exposure.
3. Accuracy Builds Confidence and Credibility
Data that aligns closely with observed conditions does more than improve technical outcomes, it builds trust with the end users. Whether you’re engaging engineers, clients, regulators, or community stakeholders, analytical transparency and agreement with ground-truth observations enhances confidence in both the process and the results. This trust is essential when asking decision-makers to plan for future risks or invest in long-term resilience.
Introducing SCOPE-ERA5: A Station-Calibrated Climate Dataset for Planning and Engineering
To address the accuracy challenges of gridded weather datasets, Degree Day has developed Station-Calibrated Outputs for Planning & Engineering (SCOPE) -ERA5, a post-processed version of ERA5 spanning 1979 to 2024. Designed specifically for planning, engineering, and climate risk applications, SCOPE-ERA5 delivers improved local accuracy by applying a peer-reviewed, multivariate bias correction algorithm trained on high-quality weather station records.
SCOPE-ERA5 enhances the accuracy of key variables relevant for infrastructure and energy, such as temperature, humidity, wind speed, and solar radiation, while preserving the observed physical relationships between them. The result is a thermodynamically consistent, multivariate dataset that more closely aligns with real-world conditions.
In the example shown earlier, SCOPE-ERA5 significantly reduces the bias in the number of days exceeding 90°F, one of many improvements in the representation of variable time series and extreme conditions.
What Does SCOPE-ERA5 Provide?
SCOPE-ERA5 offers a comprehensive suite of daily climate variables tailored for high-accuracy applications in infrastructure, energy, and climate risk analysis. The dataset is thermodynamically consistent at every time step, ensuring realistic multivariate behavior.
- Global coverage, with over 7,100 weather stations across 197 countries
- Nearly 50 years of daily data (1979–2024) calibrated to observational records
- Spatially coherent timeseries across stations
- Several variables available
A breadth of variables are provided to support a wide range of use cases, including infrastructure design, energy planning, and climate risk assessments that demand local accuracy.
- Dry-bulb temperature:
- Daily mean temperature
- Daily minimum temperature
- Daily maximum temperature
- Diurnal temperature range (DTR)
- Humidity-related variables:
- Relative humidity
- Specific humidity
- Dew point temperature
- Wet-bulb temperature
- Heat index
- Cold weather:
- Wind chill index
- Other variables:
- 10-meter surface wind speed
- Surface downwelling shortwave radiation
A technical document detailing the methodology, benchmarking, and evaluation of SCOPE-ERA5 is available here.
What are the benefits of SCOPE-ERA5?
SCOPE-ERA5 paves the way for improved assessments that reduce bias by offering a multivariate, observation-calibrated reanalysis dataset that aligns closely with real-world station data. Its internal consistency across temperature, humidity, wind speed, pressure, and solar radiation ensures that derived indices—such as heat index, wet-bulb temperature, and diurnal ranges—are locally realistic, capture extreme conditions, and adhere to thermodynamic laws. Ultimately, SCOPE-ERA5 provides practitioners with a dependable foundation for evidence-based decision-making under current and future climate conditions.
A list of use cases and ways that SCOPE-ERA5 can offer improvements is given in the table.
Use Case - Description - Benefit of SCOPE-ERA5
- Building Energy Simulation Models
Simulating HVAC performance, occupant comfort, and energy use and costs.
SCOPE-ERA5 Provides physically realistic, coincident daily values for temperature, humidity, wind speed, and solar radiation, supporting reliable building performance modeling
- Electric Grid Load Forecasting
Forecasting electricity demand and grid stress under varying weather conditions.
SCOPE-ERA5 Supplies multivariate daily inputs reflecting true local conditions, improving load accuracy estimates for extreme heat, large diurnal temperature swings, and other weather features - Renewable Energy Resource Assessment
Evaluating solar/wind generation potential and identifying low-output periods.
SCOPE-ERA5 Improve reliability and reduce inefficiencies. Delivers internally consistent, multivariate time series for assessing spatially correlated and compound risks that involve multiple variables, like solar radiation, heat and low wind.
- Public Health Planning Models
Modeling heat-related illness, exposure, and emergency response strategies
SCOPE-ERA5 Includes accurate estimates of physiologically relevant metrics like heat index and wet-bulb temperature, enabling better public health planning under extreme humid heat
- Design Days for HVAC and Building Design
Defining local climate extremes for sizing and resilience.
SCOPE-ERA5 Extends short or incomplete station records and accurately depict the tails of distributions, enabling accurate site-specific design day development
- Extreme Meteorological Year (EMY) Datasets
Stress-testing systems under anomalous weather years.
SCOPE-ERA5 Offers nearly 50 years of calibrated data to place future projected years into historical context - Trend Analysis of Multiple Variables
Assessing long-term changes in both individual variables and joint extremes.
SCOPE-ERA5 Extends short or incomplete station records, enabling robust trend analysis of individual and compound risks (e.g., dry- and wet-bulb extremes) using thermodynamically consistent, multivariate data
- Climate Model Bias Adjustment Reference
Serving as a reference for bias correction of GCM or RCM outputs
SCOPE-ERA5 Provides a physically consistent, observation-calibrated dataset to improve the accuracy of future climate projections, ensuring alignment with real-world historical baseline
Degree Day is committed to advancing climate risk research and knowledge and makes SCOPE-ERA5 data are freely available for research applications.
For more information or to request access, please contact us at info@degreeday.org
